Method and system for risk-modulated diagnosis of disease

Abstract

A method of calculating a disease assessment by analyzing a medical image, comprising (1) extracting at least one lesion feature value from the medical image; (2) extracting at least one risk feature value from the medical image; and (3) determining the disease assessment based on the at least one lesion feature value and the at least one risk feature value. The method employs lesion characterization for characterizing the lesion, and risk assessment based on the lesion's surroundings, i.e., the environment local and distal to the lesion. Computerized methods both characterize mammographic lesions and assess the breast parenchymal pattern on mammograms, resulting in improved characterization of lesions for specific subpopulations, combining the benefits of both techniques.

1. A method of calculating a disease assessment by analyzing a medical image, comprising:
 extracting at least one lesion feature value from the medical image; 
 extracting at least one risk feature value from the medical image; and 
 determining the disease assessment based on the at least one lesion feature value and the at least one risk feature value. 
 
   
   
     2. The method of  claim 1 , wherein the step of extracting the at least one lesion feature value comprises:
 locating a lesion in the medical image; 
 segmenting lesion image data corresponding to the located lesion; and 
 extracting the at least one lesion feature value from the segmented lesion image data. 
 
   
   
     3. The method of  claim 2 , wherein the step of extracting the at least one lesion feature value comprises:
 extracting at least one feature value selected from the group consisting of margin sharpness, degree of spiculation, density, homogeneity, texture, asymmetry, shape, and temporal stability of the lesion. 
 
   
   
     4. The method of  claim 2 , wherein the determining step comprises:
 determining, based on the at least one lesion feature value and the at least one risk feature value, at least one of (1) a likelihood that the lesion is malignant, (3) a stage of disease of the lesion, and (4) a likelihood that a future malignancy will develop, as the disease assessment. 
 
   
   
     5. The method of  claim 1 , wherein the step of extracting at least one risk feature value comprises:
 locating a region of interest (ROI) in the medical image corresponding to a parenchymal region; and 
 extracting the at least one risk feature value from the ROI. 
 
   
   
     6. The method of  claim 5 , wherein the step of extracting the at least one risk feature value comprises:
 extracting at least one feature value indicating a maximum gray level of the ROI, a minimum gray level of the ROI, an average gray level of the ROI, a skewness of the ROI, a coarseness of the ROI, a contrast of the ROI, a root mean square variation of a power spectrum of the ROI, and a first moment of the power spectrum of the ROI. 
 
   
   
     7. The method of  claim 1 , wherein the determining step comprises:
 calculating a quantitative measure of malignancy as the disease assessment by applying the at least one lesion feature value and the at least one risk feature value to a classifier. 
 
   
   
     8. The method of  claim 7 , wherein the calculating step comprises:
 calculating the quantitative measure of malignancy as the disease assessment by applying the at least one lesion feature value and the at least one risk feature value to a linear discriminant. 
 
   
   
     9. The method of  claim 7 , wherein the calculating step comprises:
 calculating the quantitative measure of malignancy as the disease assessment by applying the at least one lesion feature value and the at least one risk feature value to an artificial neural network. 
 
   
   
     10. The method of  claim 7 , wherein the calculating step comprises:
 training the classifier in relation to the at least one lesion feature value and the at least one risk feature value obtained from a set of previously obtained medical images based on a measure of malignancy associated with the previously obtained medical images. 
 
   
   
     11. The method of  claim 1 , wherein the determining step comprises:
 calculating a quantitative measure of risk by applying the at least one risk feature value to a first classifier; 
 calculating a quantitative measure of malignancy by applying the at least one lesion feature value to a second classifier; and 
 weighting the quantitative measure of malignancy by the quantitative measure of risk to obtain the disease assessment. 
 
   
   
     12. The method of  claim 11 , wherein:
 the step of calculating the quantitative measure of risk comprises calculating the quantitative measure of risk by applying the at least one risk feature value to a first linear discriminant; and 
 the step of calculating the quantitative measure of malignancy comprises calculating the quantitative measure of malignancy by applying the at least one lesion feature value to a second linear discriminant. 
 
   
   
     13. The method of  claim 11 , wherein:
 the step of calculating the quantitative measure of risk comprises calculating the quantitative measure of risk by applying the at least one risk feature value to a first artificial neural network; and 
 the step of calculating the quantitative measure of malignancy comprises calculating the quantitative measure of malignancy by applying the at least one lesion feature value to a second artificial neural network. 
 
   
   
     14. The method of  claim 11 , wherein the step of calculating the quantitative measure of risk comprises:
 training the first classifier in relation to the at least one risk feature value obtained from a set of previously obtained medical images based on a measure of risk associated with the previously obtained medical images. 
 
   
   
     15. The method of  claim 11 , wherein the step of calculating the quantitative measure of malignancy comprises:
 training the second classifier in relation to the at least one lesion feature value obtained from a set of previously obtained medical images based on a measure of malignancy associated with the previously obtained medical images. 
 
   
   
     16. The method of  claim 1 , wherein the determining step comprises:
 determining a quantitative measure of risk by applying the at least one risk feature value to a first classifier; 
 classifying the quantitative measure of risk as high risk if the quantitative measure of risk exceeds a predetermined threshold risk value; 
 determining the disease assessment by applying the at least one lesion feature value to a second classifier, if the quantitative measure of risk is classified as high risk; and 
 determining the disease assessment by applying the at least one lesion feature value to a third classifier, if the quantitative measure of risk is not classified as high risk. 
 
   
   
     17. The method of  claim 16 , wherein the step of calculating the quantitative measure of risk comprises:
 training the first classifier in relation to the at least one risk feature value obtained from a set of previously obtained medical images based on a measure of risk associated with the previously obtained medical images. 
 
   
   
     18. The method of  claim 16 , wherein the step of determining the disease assessment when the quantitative measure of risk is classified as high risk comprises:
 training the second classifier in relation to the at least one lesion feature value obtained from a set of previously obtained high-risk medical images based on a measure of malignancy associated with the previously obtained high-risk medical images. 
 
   
   
     19. The method of  claim 16 , wherein the step of determining the disease assessment when the quantitative measure of risk is not classified as high risk comprises:
 training the third classifier in relation to the at least one lesion feature value obtained from a set of previously obtained low-risk medical images based on a measure of malignancy associated with the previously obtained low-risk medical images. 
 
   
   
     20. The method of  claim 1 , wherein:
 the step of extracting at least one lesion feature value from the medical image comprises extracting at least one lesion feature value from a lesion extracted from a digital mammogram; and 
 the step of extracting at least one risk feature value from the medical image comprises extracting at least one risk feature value from a parenchymal region of the digital mammogram. 
 
   
   
     21. A system configured to calculate a disease assessment by analyzing a medical image by performing the steps recited in any one of  claims 1 – 20 . 
   
   
     22. A computer readable medium configured to store plural computer program instructions which, when executed by a computer, cause the computer perform the steps recited in any one of  claims 1 – 20 .